Advertisement

Severity of influenza and noninfluenza acute respiratory illness among pregnant women, 2010–2012

Open AccessPublished:August 08, 2014DOI:https://doi.org/10.1016/j.ajog.2014.08.004

      Objective

      The objective of the study was to identify characteristics of influenza illness contrasted with noninfluenza acute respiratory illness (ARI) in pregnant women.

      Study Design

      ARI among pregnant women was identified through daily surveillance during 2 influenza seasons (2010-2012). Within 8 days of illness onset, nasopharyngeal swabs were collected, and an interview was conducted for symptoms and other characteristics. A follow-up telephone interview was conducted 1-2 weeks later, and medical records were extracted. Severity of illness was evaluated by self-assessment of 12 illness symptoms, subjective ratings of overall impairment, highest reported temperature, illness duration, and medical utilization.

      Results

      Of 292 pregnant women with ARI, 100 tested positive for influenza viruses. Women with influenza illnesses reported higher symptom severity than those with noninfluenza ARI (median score, 18 vs 16 of 36; P < .05) and were more likely to report severe subjective feverishness (18% vs 5%; P < .001), myalgia (28% vs 14%; P < .005), cough (46% vs 30%; P < .01), and chills (25% vs 13%; P < .01). More influenza illnesses were associated with fever greater than 38.9°C (20% vs 5%; P < .001) and higher subjective impairment (mean score, 5.9 vs 4.8; P < .001). Differences in overall symptom severity, fever, cough, chills, early health care–seeking behavior, and impairment remained significant in multivariate models after adjusting for study site, season, age, vaccination status, and number of days since illness onset.

      Conclusion

      Influenza had a greater negative impact on pregnant women than noninfluenza ARIs, as indicated by symptom severity and greater likelihood of elevated temperature. These results highlight the importance of preventing and treating influenza illnesses in pregnant women.

      Key words

      Pregnant women are at greater risk of hospitalization and serious complications from influenza virus illness than nonpregnant women.
      • Rasmussen S.A.
      • Jamieson D.J.
      • Uyeki T.M.
      Effects of influenza on pregnant women and infants.
      Women in their third trimester
      • Hartert T.V.
      • Neuzil K.M.
      • Shintani A.K.
      • et al.
      Maternal morbidity and perinatal outcomes among pregnant women with respiratory hospitalizations during influenza season.
      • Rogers V.L.
      • Sheffield J.S.
      • Roberts S.W.
      • et al.
      Presentation of seasonal influenza A in pregnancy: 2003-2004 influenza season.
      and those with comorbid conditions
      • Cox S.
      • Posner S.F.
      • McPheeters M.
      • Jamieson D.J.
      • Kourtis A.P.
      • Meikle S.
      Hospitalizations with respiratory illness among pregnant women during influenza season.
      • Dodds L.
      • McNeil S.A.
      • Fell D.B.
      • et al.
      Impact of influenza exposure on rates of hospital admissions and physician visits because of respiratory illness among pregnant women.
      are especially vulnerable to hospitalization. Little is known about the clinical characteristics and severity of laboratory-confirmed influenza illness among nonhospitalized pregnant women or how influenza illness differs from noninfluenza acute respiratory illness (ARI) in pregnant women.
      In a previous study, we examined the relationship between vaccination status and influenza positivity.
      • Thompson M.G.
      • Li D.K.
      • Shifflett P.
      • et al.
      Effectiveness of seasonal trivalent influenza vaccine for preventing influenza virus illness among pregnant women: a population-based case-control study during the 2010-2011 and 2011-2012 influenza seasons.
      This study examined symptom severity and duration of laboratory-confirmed influenza illness among pregnant women compared with noninfluenza ARI during 2 influenza seasons. Although the number of confirmed influenza cases was limited, we conducted exploratory analyses to assess whether influenza illness severity was higher among the women with comorbid conditions or in their third trimester and whether vaccination with a seasonal influenza vaccine was associated with less severe illness among those women with vaccine failure.

      Materials and Methods

      Recruitment and eligibility

      The study methodology
      • Thompson M.
      • Williams J.
      • Naleway A.
      • et al.
      The Pregnancy and Influenza Project: design of an observational case-cohort study to evaluate influenza burden and vaccine effectiveness among pregnant women and their infants.
      and recruitment of the study participants have been described in detail previously.
      • Thompson M.G.
      • Li D.K.
      • Shifflett P.
      • et al.
      Effectiveness of seasonal trivalent influenza vaccine for preventing influenza virus illness among pregnant women: a population-based case-control study during the 2010-2011 and 2011-2012 influenza seasons.
      Eligible participants were Kaiser Permanente members who made at least 1 prenatal visit in the Portland, OR (Kaiser Permanente Northwest), or San Francisco, CA, metropolitan areas (Kaiser Permanente Northern California) during the study period, had an expected delivery date after Nov. 15, 2010, and were at least 16 years of age for Kaiser Permanente Northwest or at least 18 years of age for Kaiser Permanente Northern California. The study instruments, procedures, and written consent documents were approved by the institutional review boards at both sites.

      Surveillance

      The influenza season was defined by the regional/state health department and the Centers for Disease Control and Prevention influenza surveillance data. Thresholds for the beginning and end of the season were defined a priori at each site. During both study seasons, we identified potential ARIs using daily surveillance of electronic medical records (EMRs) for medically attended acute respiratory illness using International Classification of Diseases, ninth revision (ICD-9) codes 460-466, 480-488, and 490-491.
      The women were contacted by telephone, screened for eligibility, and asked to provide written consent for study participation. Inclusion criteria included enrollment in the health plan for the study season and completion of the enrollment interview.
      During the first season, weekly Internet- or telephone-based surveillance monitored the occurrence of non-medically attended ARI among a prospective cohort of participants at both sites.
      • Thompson M.
      • Williams J.
      • Naleway A.
      • et al.
      The Pregnancy and Influenza Project: design of an observational case-cohort study to evaluate influenza burden and vaccine effectiveness among pregnant women and their infants.
      First-season participants were encouraged to contact staff directly if they became ill; those who did not complete a weekly surveillance report received a reminder e-mail or telephone call to assess current ARI status. For both seasons, nasopharyngeal swab specimens were collected at the homes of women with self-reported cough and fever, feverishness, or chills within 8 days of illness onset.

      Participant characteristics

      Sociodemographic characteristics were assessed during an enrollment interview. Health status prior to illness was assessed with 3 measures. First, overall self-rated health was assessed using a standard question on current health (poor, fair, good, very good, or excellent) on the enrollment interview.
      • Singh-Manoux A.
      • Martikainen P.
      • Ferrie J.
      • Zins M.
      • Marmot M.
      • Goldberg M.
      What does self rated health measure? Results from the British Whitehall II and French Gazel cohort studies.
      Second, we identified high-risk comorbidities associated with an increased risk of influenza complications by the presence, during 2 or more visits over 1 year prior to conception, of the following conditions: cancer, diabetes mellitus, neurological disorders, chronic pulmonary disease, chronic cardiac disease, immunosuppressive disorders, and chronic renal disease.
      • Grohskopf L.A.
      • Shimabukuro T.T.
      • Sokolow L.Z.
      • Keitel W.A.
      • Bresee J.S.
      • Cox N.J.
      Prevention and control of seasonal influenza with vaccines: recommendations of the Advisory Committee on Immunization Practices—United States, 2013-14.
      Third, pregnancy complications, from conception to start of surveillance, were identified from EMRs, using ICD-9 codes related to adverse pregnancy outcomes. All ICD-9 codes are available upon request. Obesity was defined by body mass index, calculated using self-reported prepregnancy weight and height.

      Illness characteristics

      Illness characteristics were assessed during an illness episode interview at specimen collection and again at a follow-up telephone contact approximately 8 days later. In the first season, participants who were still ill were called again approximately 14 days after the original illness interview.
      We assessed 5 indicators of illness severity. First, participants rated the presence and severity of 12 symptoms, using a 4 point scale (0, absent; 1, mild; 2, moderate; and 3, severe). Ten participants who initially reported a cough at the screening interview described a cough as absent during their illness interview; in these instances, responses were recoded as 1 (mild). All symptom ratings were summed to form a 12 symptom severity score, ranging from 1 (a single mild symptom) to 36 (all symptoms severe), as described previously.
      • Belongia E.A.
      • Irving S.A.
      • Waring S.C.
      • et al.
      Clinical characteristics and 30-day outcomes for influenza A 2009 (H1N1), 2008-2009 (H1N1), and 2007-2008 (H3N2) infections.
      Second, participants assessed the overall subjective severity of illness from 0 (normal health) to 9 (worst possible health) and the extent of illness impairment from 0 (able to perform usual activities) to 9 (unable to do so), as described previously.
      • Henkle E IS
      • Irving S.A.
      • Naleway A.L.
      • et al.
      Comparison of laboratory-confirmed influenza and non-influenza acute respiratory illness in healthcare personnel during the 2010-2011 influenza season.
      Third, we examined febrile severity using the subjective severity of feverishness (mild, moderate, or severe) and the highest temperature recorded using any of the following: EMR vital signs (if the illness was medically attended), self-reported highest temperature at either the illness interview or 8–14 day follow-up interview(s), or measured temperature by visiting study staff. Severe fever was defined as 38.9°C, which represents the threshold for assessing teratogenic exposure in the fetus.
      Fourth, we calculated illness duration by subtracting the illness onset date (determined from screening or illness episode interview) from the date of symptom resolution as indicated in follow-up interviews. Seventy-eight women (53 noninfluenza and 25 influenza) who either were unable to recall a recovery date or had not yet recovered when the follow-up interview occurred were excluded from our analysis of illness duration.
      Fifth, we examined medical utilization and self-care as indicated by the following: (1) any or more than 1 medical visit, (2) any illness-associated hospitalization, (3) seeking health care within 2 days of onset, and (4) use of antibiotic, over-the-counter (OTC), or antiviral medications associated with the ARI as reflected in an EMR-confirmed prescription or self-reported use of a medication during the illness.

      Vaccination status

      Influenza vaccination status was documented by EMRs if available or by self-report if the participant was vaccinated outside the health plan; 14 (5%) were based on self-report. Participants who received seasonal influenza vaccine 14 days or more before illness onset were considered vaccinated.

      Laboratory methods

      Respiratory specimens were collected using nasopharyngeal swabs and stored with transport medium in cryovials at –70°C before overnight shipping on dry ice to the Marshfield Clinic Research Foundation laboratory (Marshfield, WI). Real-time reverse transcription polymerase chain reaction (RT-PCR) was performed using primer, probes, reagents, and proficiency panel provided by the Centers for Disease Control and Prevention Influenza Division (more information at http://www.cdc.gov/flu/clsis). We excluded participants whose specimens were collected more than 8 days after illness onset from our analysis.

      Statistical methods

      Six participants (2%) reported having had more than 1 ARI episode during the study period. For women testing positive for an influenza virus, we used the illness episode associated with the positive result. For influenza-negative participants, we used their first illness episode.
      We compared participant characteristics using the 2-tailed χ2 or Fisher exact test for categorical outcomes and the Student t test for continuous variables. Statistically significant (P < .05) bivariate associations between influenza and severity indicators were followed by multivariate analyses (logistic regression) to determine whether the observed association persisted after adjusting for site, season, age, days since illness onset, and vaccination status; results were expressed as adjusted odds ratios with accompanying 95% confidence intervals. Analyses were conducted using IBM SPSS (IBM, Armonk, NY) and SAS 9.3 (SAS Institute, Cary, NC) software.

      Results

      Participant and illness characteristics

      Of 1873 participants enrolled in the study, 353 pregnant women with 1 or more ARIs (107 influenza and 246 noninfluenza) were identified over 2 consecutive seasons. Seven influenza illnesses and 54 noninfluenza ARIs were excluded because illness onsets were outside the surveillance season or after delivery or because specimens were collected more than 8 days after illness onset.
      After limiting the analyses to a single illness episode per participant, we included 292 illnesses: 100 influenza-positive and 192 noninfluenza ARIs (Figure 1). All of these illnesses were medically attended, with the exception of 27 ARIs (8 influenza and 19 noninfluenza illnesses) identified as part of cohort surveillance during the first season. All 3 influenza viruses circulated in both seasons, although the median onset date of influenza illnesses was 3 weeks later in the second season (Feb. 22, 2011, vs March 14, 2012; Appendix; Supplemental Figure).
      Figure thumbnail gr1
      Figure 1Flow chart showing enrollment and exclusions for each study season
      ARI, acute respiratory illness; S1, season 1 (2010-2011); S2, season 2 (2011-2012).
      Sokolow. Influenza severity in pregnant women. Am J Obstet Gynecol 2015.
      Influenza positivity differed by days since illness onset when the specimen was collected (P < .05); this difference was due to the relatively high influenza positivity on day 1 since illness onset (65%) compared with a varying range of positivity noted on days 2-8 (26-44%). Combining both seasons, 45 women had RT-PCR–confirmed A(H1N1)pdm09 virus infections, 33 had A(H3N2), 1 had an unsubtyped A virus, and 21 had influenza B virus infections.
      Participant characteristics are presented in Table 1. About half of participants (54%) were white non-Hispanics, and 24% were Hispanics. At the time of illness onset, the mean age was 30.7 years, and most women were in their second (44%) or third (41%) trimester.
      Table 1Sociodemographic characteristics of pregnant women with influenza vs noninfluenza ARIs
      CharacteristicsNoninfluenza ARI (n = 192)Influenza (n = 100)P valueTotal (n = 292)
      Study site
       Kaiser Permanente Northern California111(58)69(69)NS180(62)
       Kaiser Permanente Northwest81(42)31(31)112(38)
      Study season
       2010-2011115(60)53(53)NS168(58)
       2011-201277(40)47(47)124(42)
      Age, y, at illness onset
       Mean ± SD30.9±5.830.5±5.8NS30.7±5.8
      Race/ethnicity
       Hispanic45(23)25(25)NS70(24)
       White, non-Hispanic110(57)48(48)158(54)
       Asian, non-Hispanic14(7)13(13)27(9)
       Other, non-Hispanic23(12)14(14)37(13)
      Trimester at illness onset date
       First (7-13 wks)31(16)13(13)NS44(15)
       Second (14-26 wks)85(44)44(44)129(44)
       Third (27-42 wks)76(40)43(43)119(41)
      Recent health history, mean days ± SD
       For how many days during past 30 days was…
      Your physical health not good?7.1±9.72.0±4.9< .0015.2±8.6
      Your mental health not good?4.8±8.83.0±5.4NS4.1±7.7
       Self-rated health status, at time of enrollment
      Very good or excellent116(61)68(68)NS184(63)
      Good60(31)27(27)87(30)
      Fair or poor16(8)5(5)21(7)
      High-risk comorbidities
      High-risk comorbidities, identified during 1 year prior to conception, for which influenza vaccination is recommended by the ACIP, including cancer, diabetes mellitus, neurological disorders, chronic pulmonary disease, chronic cardiac disease, immunosuppressive disorders, chronic renal disease (codes available upon request)
      At least 157(30)17(17)< .0574(25)
      >123(12)2(2)< .00525(9)
       History of asthma
      Any asthma ICD-9 code (493) in the electronic medical record within the prior year to conception
      27(14)2(2)< .00529(10)
       Obese (BMI >30 kg/m2)
      BMI was calculated using self-reported prepregnancy weight and height
      59(31)15(15)< .0574(25)
       High-risk pregnancy
      Pregnancy complication was indicated by a medical visit associated with a subset of ICD-9 codes related to adverse pregnancy outcomes from conception through the end of the vaccine campaign for the study year (codes available upon request)
      41(21)24(24)NS65(22)
      Influenza vaccination status
       Current season
      Current season vaccinated defined as receipt of vaccine at least 14 days prior to illness onset.
      112(58)42(42)< .01154(53)
       Current season and at least 1 comorbidity40(21)9(9)< .0149(17)
      Data are counts (column percentage), unless otherwise indicated. P values are from Pearson χ2, Fisher exact test, or Student t test.
      ACIP, Advisory Committee on Immunization Practices; ARI, acute respiratory illness; BMI, body mass index; ICD-9, International Classification of Diseases, ninth revision; NS, not statistically significant.
      Sokolow. Influenza severity in pregnant women. Am J Obstet Gynecol 2015.
      a High-risk comorbidities, identified during 1 year prior to conception, for which influenza vaccination is recommended by the ACIP, including cancer, diabetes mellitus, neurological disorders, chronic pulmonary disease, chronic cardiac disease, immunosuppressive disorders, chronic renal disease (codes available upon request)
      b Any asthma ICD-9 code (493) in the electronic medical record within the prior year to conception
      c BMI was calculated using self-reported prepregnancy weight and height
      d Pregnancy complication was indicated by a medical visit associated with a subset of ICD-9 codes related to adverse pregnancy outcomes from conception through the end of the vaccine campaign for the study year (codes available upon request)
      e Current season vaccinated defined as receipt of vaccine at least 14 days prior to illness onset.
      Women with influenza illnesses were similar to those with noninfluenza ARIs on most sociodemographic and health indicators, with 2 exceptions: women with influenza were less likely to have a preexisting comorbidity (17% vs 30%; P < .05), including asthma (2% vs 14%; P < .005) or obesity before pregnancy (15% vs 31%; P < .05), and women with influenza reported fewer days of poor physical health in the prior 30 days (2.0 vs 7.1 days; P < .001). The lower influenza positivity associated with high-risk conditions and poor health remained statistically significant after adjusting for the higher influenza vaccination rate for those in poor health (values of P < .05).

      Comparisons of influenza and noninfluenza ARI severity

      Small but significant differences in symptom severity and illness impairment were observed between influenza and noninfluenza ARI. Specifically, the median summed subjective severity of 12 symptoms was 18 for influenza ARI vs 16 (of 36) for noninfluenza ARI (P < .05; Table 2). Women with influenza rated their illnesses as causing greater impairment to their activities than women with noninfluenza ARI (5.9 vs 4.8 on a 9 point scale; P < .001).
      Table 2Clinical characteristics associated with influenza vs noninfluenza ARIs among pregnant women
      CharacteristicsNoninfluenza ARI (n = 192)Influenza (n = 100)P value
      Illness and specimen collection timing
       Days from onset to swab, mean days ± SD4.7±1.84.2±2.0< .05
      Subjective symptom severity and impact
       12-symptom severity score, median/IQR
      Summed subjective severity of 12 symptoms: cough, fatigue, feverishness, congestion, headache, myalgia, sore throat, chills, wheezing, ear pain, nausea, and vomiting. Each symptom was rated as absent (0), mild (1), moderate (2), or severe (3), except for cough, which did not have a response option (0)
      1612-201814-20< .05
       Overall subjective severity, mean score ± SD
      Participants were asked to rate their health during their illness (0, normal health; 9, worse possible health)
      4.7±1.84.3±2.0NS
       Subjective impairment, mean score ± SD
      Participants were asked to rate their ability to perform activities during their illness (0, able to perform usual activities; 9, unable to perform usual activities; severe was defined as an impairment rate of ≥6)
      4.8±2.35.9±1.9< .001
       Illness duration, mean days ± SD
      Mean illness duration was calculated using only the response from those participants who could recall the date their illness ended (noninfluenza ARI [n = 139] and influenza [n = 75])
      10.3±4.89.3±4.2NS
      Fever severity
       Any fever, ≥37.8°C47(24)63(63)< .001
       Highly elevated fever of ≥38.9°C9(5)20(20)< .001
       Highest recorded temperature, mean °C ± SD37.2±0.838.0±1.0< .001
      Medical utilization during illness
       Any medical visit for ARI
      Medically attended acute respiratory illness visits that occurred between onset and recovery date with at least 1 of these ICD-9 codes: 460-466, 480-488 or 490-491
      156(81)88(88)NS
       Two or more medical visits for ARI76(40)42(42)NS
       Early care-seeking behavior
      Medically attended 2 days or less following onset
      68(35)53(53)< .001
       First medical encounter, mean days ± SD
      Days between illness onset and first medical visit (includes only those people who had medical encounters).
      3.0±2.12.4±2.2< .05
       Emergency department visits15(8)12(12)NS
       Illness-associated hospitalization7(4)5(4)NS
       Use of medications
      Antibiotic prescription
      Prescribed30(16)11(11)NS
      Self-reported use21(11)9(9)NS
      Antiviral prescription
      Prescribed2(1)8(8)NS
      Self-reported use6(3)10(10)NS
      OTC medications, self-reported use155(81)92(92)< .05
      Data are counts (column percentage), unless otherwise indicated. P values are from Pearson χ2, Fisher exact, or Student t tests.
      ARI, acute respiratory illness; ICD-9, International Classification of Diseases, ninth revision; IQR, interquartile range; NS, not statistically significant; OTC, over-the-counter medications.
      Sokolow. Influenza severity in pregnant women. Am J Obstet Gynecol 2015.
      a Summed subjective severity of 12 symptoms: cough, fatigue, feverishness, congestion, headache, myalgia, sore throat, chills, wheezing, ear pain, nausea, and vomiting. Each symptom was rated as absent (0), mild (1), moderate (2), or severe (3), except for cough, which did not have a response option (0)
      b Participants were asked to rate their health during their illness (0, normal health; 9, worse possible health)
      c Participants were asked to rate their ability to perform activities during their illness (0, able to perform usual activities; 9, unable to perform usual activities; severe was defined as an impairment rate of ≥6)
      d Mean illness duration was calculated using only the response from those participants who could recall the date their illness ended (noninfluenza ARI [n = 139] and influenza [n = 75])
      e Medically attended acute respiratory illness visits that occurred between onset and recovery date with at least 1 of these ICD-9 codes: 460-466, 480-488 or 490-491
      f Medically attended 2 days or less following onset
      g Days between illness onset and first medical visit (includes only those people who had medical encounters).
      In terms of specific symptoms, women with influenza were more likely to report severe subjective feverishness (18% vs 5%; P < .001), myalgia (28% vs 14%, P < .005), cough (46% vs 30%, P < .01), and chills (25% vs 13%; P < .01; Table 3). On average, women with influenza described their cough, nasal congestion, and fatigue as severe; however, only cough was statistically more severe for those with influenza (Figure 2); myalgia, fever, and chills were more severe among those with influenza, but the mean severity rating for each of these symptoms was moderate.
      Table 3Self-reported symptom severity measures associated with influenza vs noninfluenza ARIs among pregnant women
      CharacteristicsPercentage self-reported as severe
      Noninfluenza ARI (n = 192)Influenza (n = 100)P value
      Upper respiratory symptoms
       Nasal congestion90(47)42(42)NS
       Sore throat65(34)23(23)NS
       Ear pain17(9)11(11)NS
      Lower respiratory symptoms
       Cough
      Ten participants reported cough as absent during their illness: in these instances, their responses were recoded as 1 (mild) because a cough was present at the time they were deemed eligible to participate in the study.
      57(30)46(46)< .01
       Wheezing11(6)4(4)NS
      Systemic symptoms
       Fatigue66(34)41(41)NS
       Myalgia26(14)28(28)< .005
       Feverishness9(5)18(18)< .001
       Chills25(13)25(25)< .01
       Headache48(25)18(18)NS
      Gastrointestinal symptoms
       Vomiting15(8)5(5)NS
       Nausea18(9)5(5)NS
      Data are counts (column percentage), unless otherwise indicated. P values are from Pearson χ2, Fisher exact, or Student t tests.
      ARI, acute respiratory illness; NS, not statistically significant.
      Sokolow. Influenza severity in pregnant women. Am J Obstet Gynecol 2015.
      a Ten participants reported cough as absent during their illness: in these instances, their responses were recoded as 1 (mild) because a cough was present at the time they were deemed eligible to participate in the study.
      Figure thumbnail gr2
      Figure 2Mean response to self-reported severity ranking of 12 symptoms
      Mean response to self-reported severity ranking of 12 symptoms, with 95% confidence intervals. Symptoms are sorted by mean severity score: severe, ≥2; moderate, 1–1.9; mild, >1. Vertical line with solid circle indicates influenza illnesses; vertical line with open circle indicates noninfluenza ARI. Asterisk indicates symptoms with statistically different responses between influenza and noninfluenza ARI (P < .01).
      ARI, acute respiratory illness.
      Sokolow. Influenza severity in pregnant women. Am J Obstet Gynecol 2015.
      More women with influenza had an elevated (≥37.8°C) or highly elevated (≥38.9°C) recorded temperature (63% and 20%, respectively) compared with women with noninfluenza ARI (24% and 5%, respectively; all P < .001 (Table 2). Illness duration was similar for women with influenza and noninfluenza ARI (mean, 9.3 vs 10.3, respectively).
      Women with influenza sought medical care sooner: 53% within 2 days of illness onset, compared with 35% among women with noninfluenza ARI (P < .001; Table 2). Slightly more women with influenza reported using OTC medications (92% vs 81%; P < .05), but few in either group were prescribed or used antibiotic or antiviral medications. In fact, few women with influenza were prescribed antiviral medications (8%), and there was no difference in the likelihood of either group to be prescribed antibiotic medications (11% vs 16%).
      In multivariate analyses adjusting for site, season, age, vaccination status, and days since illness onset, several variables showed significant differences between influenza and noninfluenza ARI (Table 4). The adjusted odds of having a temperature greater than 38.9°C or severe subjective feverishness was greater than 4-fold higher among women with influenza. The adjusted odds were 2 times higher for activity impairment, early care-seeking behavior, a higher 12 symptom severity score, severe subjective cough, and chills.
      Table 4Results of logistic regression of factors associated with influenza vs noninfluenza acute respiratory illnesses among 292 pregnant women, 2 sites, 2010-2012
      CharacteristicsaOR [95% CI]
      Adjusted for site, season, age, vaccination status, and days from illness onset to influenza testing.
      P value
      Site (referent, KPNC)1.66 [0.98–2.82].062
      Season (referent, 2010-2011)0.76 [0.45–1.27].288
      Age at illness onset (referent, <31 y)0.78 [0.48–1.27].317
      Vaccination status in current season (referent, unvaccinated)0.58 [0.35–0.96].034
      Influenza test 6 days or longer after onset (referent, <6 d)0.75 [0.44–1.28].289
      Highest recorded temperature ≥38.9°C (referent, <38.9°C)4.51 [1.92–10.56]< .0001
      Activities greatly impaired (referent, not greatly impaired)2.43 [1.42–4.18].001
      Early care-seeking behavior (referent, visit >2 d after onset)2.28 [1.36–3.82].002
      12 symptom severity score (referent, not severe)2.06 [1.10–3.87].025
      Subjective feverishness severe (referent, not severe)4.08 [1.71–9.70].002
      Subjective cough severe (referent, not severe)2.05 [1.21–3.46].007
      Subjective chills severe (referent, not severe)2.01 [1.06–3.80].032
      aOR, adjusted odds ratio; CI, confidence interval; KPNC, Kaiser Permanente Northern California.
      Sokolow. Influenza severity in pregnant women. Am J Obstet Gynecol 2015.
      a Adjusted for site, season, age, vaccination status, and days from illness onset to influenza testing.

      Impact of vaccination status, comorbidities, and trimester on influenza illness severity

      Among 100 women with influenza illnesses, illness severity did not differ by trimester, receipt of influenza vaccine, or the presence of comorbidities (Supplemental Table). Within the same group, a temperature of 38.9°C or greater was more common among those vaccinated vs unvaccinated (40% vs 17%, P < .01), but the severity of subjective symptoms did not differ among those vaccinated vs unvaccinated (data not shown).

      Secondary analyses

      Despite the small number of influenza cases by influenza (sub)type, we conducted exploratory analyses to compare severity indicators by influenza type and subtype. The mean symptom severity score was higher for influenza B illnesses than either influenza A(H3N2) (19.7 vs 17.4, P < .05) or influenza A(H1N1)pdm09 (19.7 vs 17.2, P < .05; Figure 3). There were no significant differences in illness duration (means, 9.4 days for influenza B, 8.0 for A[H3N2], and 8.5 for A[H1N1]pmd09 virus infections).
      Figure thumbnail gr3
      Figure 3Mean sum of 12 self-reported symptoms by influenza virus type/subtype
      Mean sum of 12 self-reported symptoms by influenza virus type/subtype compared with noninfluenza illnesses, with 95% confidence intervals shown. Individual summed scores ranged from 6 to 29. One case with an unsubtypable influenza illness has been excluded from this figure. Mean severity scores for influenza B and noninfluenza are statistically different. Symptoms include cough, fatigue, feverishness, congestion, headache, myalgia, sore throat, chills, wheezing, ear pain, nausea, and vomiting. Each symptom was rated as absent (0), mild (1), moderate (2), or severe (3), except for cough, which did not have a mild response option.
      Sokolow. Influenza severity in pregnant women. Am J Obstet Gynecol 2015.
      In a sensitivity analysis, we limited participants to those with illness onset of 4 days or less from the home visit and respiratory specimen collection and noted no difference in the direction or pattern of effects. We also excluded non–medically attended illnesses and found it also did not change the main findings.

      Comment

      We found that pregnant women with influenza illnesses described some symptoms as more severe (especially cough, myalgia, fever, and chills), reported greater subjective impairment, and were more likely to seek care earlier and to have highly elevated temperatures. Among the women infected with influenza, influenza vaccination was not associated with less severe illness.
      We found a median symptom severity score of 18 (of 36) for influenza illnesses in pregnant women, which was within the range that Belongia et al
      • Belongia E.A.
      • Irving S.A.
      • Waring S.C.
      • et al.
      Clinical characteristics and 30-day outcomes for influenza A 2009 (H1N1), 2008-2009 (H1N1), and 2007-2008 (H3N2) infections.
      reported using for the same measure for nonhospitalized adults infected with 3 different influenza A strains, using the same scale. Women with influenza were more likely to seek medical care within 2 days of onset, which suggests that severe symptoms present more rapidly after onset. More than half of episodes (53%) were associated with an medical encounter within the first 2 days of onset, similar to previously reported estimates of 45-60% for influenza A among community-dwelling adults and children.
      • Irving S.A.
      • Patel D.C.
      • Kieke B.A.
      • et al.
      Comparison of clinical features and outcomes of medically attended influenza A and influenza B in a defined population over four seasons: 2004-2005 through 2007-2008.
      Women with influenza rated the impairment caused by their illness as greater than those with noninfluenza ARI, similar to a recent study with health care workers using the same measures.
      • Henkle E IS
      • Irving S.A.
      • Naleway A.L.
      • et al.
      Comparison of laboratory-confirmed influenza and non-influenza acute respiratory illness in healthcare personnel during the 2010-2011 influenza season.
      This study has several strengths. We directly compared the epidemiological and clinical characteristics of influenza and noninfluenza ARI in the same population of pregnant women using identical enrollment, laboratory diagnostics, and follow-up methods. We used EMRs to conduct surveillance to identify women eligible for testing and to determine vaccination status and comorbidities, thereby reducing recall biases, and we used RT-PCR testing, collection of detailed data regarding clinical and subjective manifestations of illness, and quick response time between illness detection, testing for influenza, and then interviewing participants about their illness.
      Our study also has several limitations. The highest fever reached during the illness was an important measure of severity; however, we had to rely on the highest reported temperature, which may understate the true severity of fever if women are not consistently and repeatedly taking their temperature. We were also unable to determine the duration of fever. Our focus on febrile ARI with cough prevented us from considering mild or atypical manifestations of influenza illness. A report of 2 seasons may reflect the clinical manifestations of circulating influenza viruses during those seasons and may not be generalizable to other strains or seasons. The small number of influenza illnesses available for analysis limited our power to examine differences among influenza (sub)types, vaccination status, trimester at onset, or more serious illness outcomes (eg, hospitalization). We were unable to disentangle the contributions of host, agent, and environment in influencing the risk of infection or manifestations of disease.
      With a larger sample size, it would be useful to explore whether a combination of symptoms and severity could effectively rule in or rule out influenza prior to receiving test results. The relatively small number of influenza and noninfluenza illnesses examined in this study limited the precision of our findings; future studies with more precise estimates of the positive and negative predictive value of illness symptoms may facilitate potential clinical applications.
      • Eubell M.H.
      • Afonso A.M.
      • Gonzales R.
      • et al.
      Development and validation of a clinical decision reule for the diagnosis of influenza.
      Medication usage could have affected severity and clinical outcomes. Use of antiviral and antibiotics was similar across groups; however, more influenza-positive women took OTC medications, which could have masked severity. Some self-reported data (eg, symptom severity, use of OTC medications, smoking status, and self-reported vaccine status for the 14 women vaccinated outside the Kaiser Permanente system [5%]) could not be verified.

      Comparisons of influenza and noninfluenza ARI severity

      This study indicates that influenza and noninfluenza ARIs in pregnant women are distinguishable based on subjective and EMR-based severity measurements. These differences persisted after adjusting for site, season, age, vaccination status, and days since illness onset. Our logistic regression analyses revealed a strong association between an influenza-positive illness and severe cough and/or fever. Overall, our findings generally are consistent with prior studies of influenza illnesses.
      • Belongia E.A.
      • Irving S.A.
      • Waring S.C.
      • et al.
      Clinical characteristics and 30-day outcomes for influenza A 2009 (H1N1), 2008-2009 (H1N1), and 2007-2008 (H3N2) infections.
      • Henkle E IS
      • Irving S.A.
      • Naleway A.L.
      • et al.
      Comparison of laboratory-confirmed influenza and non-influenza acute respiratory illness in healthcare personnel during the 2010-2011 influenza season.
      • Irving S.A.
      • Patel D.C.
      • Kieke B.A.
      • et al.
      Comparison of clinical features and outcomes of medically attended influenza A and influenza B in a defined population over four seasons: 2004-2005 through 2007-2008.
      Although most influenza and noninfluenza ARI illnesses involved fever, feverishness, or chills, records of highly elevated temperature (>38.9°C) were much more common among influenza (20% vs 5% noninfluenza ARI). High fever is a cause for concern because a temperature elevation of 1.9°C over normal (37.0°C) is the threshold for assessing teratogenic risk (eg, neural tube defects).
      • Chambers C.D.
      • Johnson K.A.
      • Dick L.M.
      • Felix R.J.
      • Jones K.L.
      Maternal fever and birth outcome: a prospective study.
      • Edwards M.J.
      Review: hyperthermia and fever during pregnancy.
      There is little information on the effects of higher temperature elevations for short exposures or more modest elevations for longer periods of time, although both situations can yield a similar thermal dose (ie, length and intensity of fever). We were unable to measure fever duration; however, 20% of influenza illnesses reached the temperature risk threshold. After adjusting for site, season, age, vaccination status, and days since illness onset, we found the odds of reaching or exceeding this risk threshold were greater than 4.5-fold higher among influenza-positive women.
      Although antiviral medications may reduce complications of influenza illness and shorten symptom duration,
      • Hernan M.A.
      • Lipsitch M.
      Oseltamivir and risk of lower respiratory tract complications in patients with flu symptoms: a meta-analysis of eleven randomized clinical trials.
      • Hsu J.
      • Santesso N.
      • Mustafa R.
      • et al.
      Antivirals for treatment of influenza: a systematic review and meta-analysis of observational studies.
      • Jefferson T.
      • Jones M.A.
      • Doshi P.
      • et al.
      Neuraminidase inhibitors for preventing and treating influenza in healthy adults and children.
      few women in our study with influenza illnesses (10%) were prescribed antivirals; this is a somewhat lower rate of use than was reported previously (15-22%) for nonpregnant adults.
      • Belongia E.A.
      • Irving S.A.
      • Waring S.C.
      • et al.
      Clinical characteristics and 30-day outcomes for influenza A 2009 (H1N1), 2008-2009 (H1N1), and 2007-2008 (H3N2) infections.
      • Irving S.A.
      • Patel D.C.
      • Kieke B.A.
      • et al.
      Comparison of clinical features and outcomes of medically attended influenza A and influenza B in a defined population over four seasons: 2004-2005 through 2007-2008.
      The reasons for this relatively low level of antiviral use probably include the fact that only 36% of women in our study sought medical care within 2 days after illness onset and within the ideal window for antiviral treatment. The lower use of antiviral medications in some integrated care systems
      • Havers F.
      • Thaker S.
      • Clippard J.R.
      • et al.
      Use of influenza antiviral agents by ambulatory care clinicians during the 2012-13 influenza season.
      and the hesitancy of pregnant women to take medications in general
      • Nordang H.
      • Koren G.
      • Einarson A.
      Pregnant women’s beliefs about medication—a study among 866 Norwegian women.
      may have also contributed.

      Impact of trimester and vaccination status on influenza illness severity

      We found no evidence that influenza vaccination was associated with lower subjective symptom severity. Vaccinated women were more likely to have a high fever of 38.9°C or greater (40% vs 17%; P < .01); we cannot explain this finding. Power for analyzing the impact of vaccination on other illness severity characteristics was low because of the small number of influenza illnesses when separated by vaccination status (42 vaccinated and 58 unvaccinated).
      Patients were diagnosed in all trimesters of pregnancy, and we found little evidence that influenza illnesses in third-trimester pregnant women were more severe than those that occurred earlier. Women with third-trimester illnesses did have fewer severe nausea symptoms, but it is difficult to differentiate typical early pregnancy nausea from nausea caused by influenza.

      Impact of influenza type/subtype on illness severity

      Women infected with influenza B viruses were more likely to have severe symptoms than either influenza A subtype (mean severity score, 19.7 vs 17.4 for influenza A[H3N2] and 17.2 for influenza A[H1N1]pdm09). Although the severity of influenza B virus infections among adults is debated,
      • McCullers J.A.
      • Hayden F.G.
      Fatal influenza B infections: time to reexamine influenza research priorities.
      this finding differs from some published studies. For example, Irving et al
      • Irving S.A.
      • Patel D.C.
      • Kieke B.A.
      • et al.
      Comparison of clinical features and outcomes of medically attended influenza A and influenza B in a defined population over four seasons: 2004-2005 through 2007-2008.
      found that over 4 seasons in 2007–2011, influenza B illnesses were milder than influenza A. The discrepancy could reflect seasonal variation in the virulence of influenza B viruses.
      • McCullers J.A.
      • Hayden F.G.
      Fatal influenza B infections: time to reexamine influenza research priorities.

      Conclusion

      Influenza had a greater negative impact on pregnant women than noninfluenza ARIs. This finding adds to a growing literature indicating that illness caused by influenza virus infection, even among those not requiring hospitalization, is moderately more severe than illnesses caused by other respiratory pathogens and reinforces the importance of influenza prevention and control efforts among pregnant women and women planning to become pregnant in the next 12 months.
      • Grohskopf L.A.
      • Shimabukuro T.T.
      • Sokolow L.Z.
      • Keitel W.A.
      • Bresee J.S.
      • Cox N.J.
      Prevention and control of seasonal influenza with vaccines: recommendations of the Advisory Committee on Immunization Practices—United States, 2013-14.

      Acknowledgments

      Members of the Pregnancy and Influenza Project Workgroup, in addition to the authors, included the following: Susan Chu, Janet Cragan, Anne McIntyre, Julie Villanueva, and Jennifer L. Williams. We thank the following people for their important contributions: Abt Associates: Sarah Ball, Sam Bozeman, Beth Brenna, Rebecca Fink Leigh Mathias, and Rachael Wendlandt; Kaiser Permanente Northern California: Clio Bernhardson-Massolo, Giulia Chillemi, Monica Flores, Shannon Growl, Crystal Warren; Kaiser Permanente Northwest: Tia Kauffman, Samantha Kurosky, and Nancy Siegel; Centers for Disease Control and Prevention: Jufu Chen, Po-Yung Cheng, Samuel Peasah, Jeffrey Sokolow, Craig Steffen, and Hong (Holly) Zhou.

      Appendix

      Figure thumbnail fx1
      Supplemental FigurePregnant women with febrile acute respiratory illness
      Number of pregnant women with febrile acute respiratory illness (ARI) who tested positive or negative for influenza by week of testing during 2010-2011 and 2011-2012 seasons.
      Sokolow. Influenza severity in pregnant women. Am J Obstet Gynecol 2015.
      Supplemental TableThe influence of selected factors on self-reported symptom severity measures among 100 pregnant women with laboratory-confirmed influenza during 2010-2012 influenza season, 2 sites
      SeverityIllness severity
      Summed subjective severity of 12 symptoms were defined as severe if score was 22-29
      Total (n = 100)
      Mild or moderate (n = 54)Severe (n = 26)P value
      Trimester at illness onset
       First or second trimester42(74)15(26)NS57(100)
       Third trimester32(74)11(26)43(100)
      Current season vaccination status
      Current-season vaccination status defined as receipt of vaccine at least 14 days prior to illness onset.
       Vaccinated43(74)15(26)NS58(100)
       Not vaccinated31(74)11(26)42(100)
      Comorbidities
       None60(72)23(28)NS83(100)
       One or more14(82)3(18)17(100)
      Data are counts (row percentage). P values are from Pearson χ2 or Fisher exact test.
      NS, not statistically significant.
      Sokolow. Influenza severity in pregnant women. Am J Obstet Gynecol 2015.
      a Summed subjective severity of 12 symptoms were defined as severe if score was 22-29
      b Current-season vaccination status defined as receipt of vaccine at least 14 days prior to illness onset.

      References

        • Rasmussen S.A.
        • Jamieson D.J.
        • Uyeki T.M.
        Effects of influenza on pregnant women and infants.
        Am J Obstet Gynecol. 2012; 207: S3-S8
        • Hartert T.V.
        • Neuzil K.M.
        • Shintani A.K.
        • et al.
        Maternal morbidity and perinatal outcomes among pregnant women with respiratory hospitalizations during influenza season.
        Am J Obstet Gynecol. 2003; 189: 1705-1712
        • Rogers V.L.
        • Sheffield J.S.
        • Roberts S.W.
        • et al.
        Presentation of seasonal influenza A in pregnancy: 2003-2004 influenza season.
        Obstet Gynecol. 2010; 115: 924-929
        • Cox S.
        • Posner S.F.
        • McPheeters M.
        • Jamieson D.J.
        • Kourtis A.P.
        • Meikle S.
        Hospitalizations with respiratory illness among pregnant women during influenza season.
        Obstet Gynecol. 2006; 107: 1315-1322
        • Dodds L.
        • McNeil S.A.
        • Fell D.B.
        • et al.
        Impact of influenza exposure on rates of hospital admissions and physician visits because of respiratory illness among pregnant women.
        CMAJ. 2007; 176: 463-468
        • Thompson M.G.
        • Li D.K.
        • Shifflett P.
        • et al.
        Effectiveness of seasonal trivalent influenza vaccine for preventing influenza virus illness among pregnant women: a population-based case-control study during the 2010-2011 and 2011-2012 influenza seasons.
        Clin Infect Dis. 2014; 58: 449-457
        • Thompson M.
        • Williams J.
        • Naleway A.
        • et al.
        The Pregnancy and Influenza Project: design of an observational case-cohort study to evaluate influenza burden and vaccine effectiveness among pregnant women and their infants.
        Am J Obstet Gynecol. 2011; 204: S69-S76
        • Singh-Manoux A.
        • Martikainen P.
        • Ferrie J.
        • Zins M.
        • Marmot M.
        • Goldberg M.
        What does self rated health measure? Results from the British Whitehall II and French Gazel cohort studies.
        J Epidemiol Community Health. 2006; 60: 364-372
        • Grohskopf L.A.
        • Shimabukuro T.T.
        • Sokolow L.Z.
        • Keitel W.A.
        • Bresee J.S.
        • Cox N.J.
        Prevention and control of seasonal influenza with vaccines: recommendations of the Advisory Committee on Immunization Practices—United States, 2013-14.
        MMWR Morb Mortal Wkly Rep. 2013; 62: 1-43
        • Belongia E.A.
        • Irving S.A.
        • Waring S.C.
        • et al.
        Clinical characteristics and 30-day outcomes for influenza A 2009 (H1N1), 2008-2009 (H1N1), and 2007-2008 (H3N2) infections.
        JAMA. 2010; 304: 1091-1098
        • Henkle E IS
        • Irving S.A.
        • Naleway A.L.
        • et al.
        Comparison of laboratory-confirmed influenza and non-influenza acute respiratory illness in healthcare personnel during the 2010-2011 influenza season.
        Infect Control Hosp Epidemiol. 2014; 35: 538-546
        • Irving S.A.
        • Patel D.C.
        • Kieke B.A.
        • et al.
        Comparison of clinical features and outcomes of medically attended influenza A and influenza B in a defined population over four seasons: 2004-2005 through 2007-2008.
        Influenza Other Respir Viruses. 2012; 6: 37-43
        • Eubell M.H.
        • Afonso A.M.
        • Gonzales R.
        • et al.
        Development and validation of a clinical decision reule for the diagnosis of influenza.
        J Am Board Fam Med. 2012; 25: 55-62
        • Chambers C.D.
        • Johnson K.A.
        • Dick L.M.
        • Felix R.J.
        • Jones K.L.
        Maternal fever and birth outcome: a prospective study.
        Teratology. 1998; 58: 251-257
        • Edwards M.J.
        Review: hyperthermia and fever during pregnancy.
        Birth Defects Res A Clin Mol Terat. 2006; 76: 507-516
        • Hernan M.A.
        • Lipsitch M.
        Oseltamivir and risk of lower respiratory tract complications in patients with flu symptoms: a meta-analysis of eleven randomized clinical trials.
        Clin Infect Dis. 2011; 53: 277-279
        • Hsu J.
        • Santesso N.
        • Mustafa R.
        • et al.
        Antivirals for treatment of influenza: a systematic review and meta-analysis of observational studies.
        Ann Intern Med. 2012; 156: 512-524
        • Jefferson T.
        • Jones M.A.
        • Doshi P.
        • et al.
        Neuraminidase inhibitors for preventing and treating influenza in healthy adults and children.
        Cochrane Database Syst Rev. 2012; 1: CD008965
        • Havers F.
        • Thaker S.
        • Clippard J.R.
        • et al.
        Use of influenza antiviral agents by ambulatory care clinicians during the 2012-13 influenza season.
        Clin Infect Dis. 2014; 59: 774-782
        • Nordang H.
        • Koren G.
        • Einarson A.
        Pregnant women’s beliefs about medication—a study among 866 Norwegian women.
        Ann Pharmacother. 2010; 44: 1478-1484
        • McCullers J.A.
        • Hayden F.G.
        Fatal influenza B infections: time to reexamine influenza research priorities.
        J Infect Dis. 2012; 205: 870-872